Slurm users reach a node only through a job of theirs

llmgw's Slurm users log in to the controller freely and to another Slurm
machine only while a job of theirs runs there: the agent installs
pam_slurm_adopt and puts it in sshd's account stack for the users it made
alone, so root and the machine's own users are never locked out. Their
session is put in the job's extern step (PrologFlags=Contain), and
pam_systemd, which would move it out, is skipped for them by taking the
session file into sshd's stack with a guard in front of it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

CHANGELOG.md

@@ -60,7 +60,11 @@ APIs change between minors without a deprecation period.
   llmgw` (`ctl slurm users on`) makes each active account a Linux user on
   every Slurm machine, the same login and UID (from 200000, never reused)
   everywhere, its home on the shared path; people add SSH keys in the
-  portal's Profile (`ctl slurm users keys` for an operator).
+  portal's Profile (`ctl slurm users keys` for an operator). They log in to
+  the controller; to another Slurm machine only while a job of theirs runs
+  there, their session put in the job (pam_slurm_adopt), so nobody uses a
+  node's cards outside Slurm, nor on a node llmgw serves from. Root and the
+  machines' own users are not held to it.
 
 ### Changed